Collaboration in Supply Chains: With and Without Trust

John Paul MacDuffie Wharton School, U. of Pennsylvania
3105 Steinberg-Dietrich Hall Philadelphia, PA 19104 215-898-2588

Susan Helper Case Western Reserve University
272 Peter B. Lewis Building Cleveland, OH 44106 216-368-5541 Version: June 14, 2005 Forthcoming in Collaborative Community Charles Heckscher and Paul S. Adler, editors, Oxford University Press, 2005

The authors are grateful for comments received from the editors, colleagues from the International Motor Vehicle Program (IMVP), particularly Mari Sako, Takahiro Fujimoto, Akira Takeishi, Sebastian Fixson, and Frits Pil; to participants in the Jones Center seminar, Wharton School; the ICOS seminar, U. of Michigan, and to the individuals at various OEMs and suppliers whom we interviewed. We are also grateful to IMVP for providing the funding for this research.


Collaboration in Supply Chains: With and Without Trust
The growth of collaboration has occurred over the last few decades not only within firms but also across firms, through elaboration of complex supply-chain and alliance relationships. The automobile industry has long been used as an exemplar of important economic phenomena involving supply chains. For most of the first two-thirds of the 20th century, mass production was dominant and the automotive industry was highly vertically integrated. During this period, economists dating back to Coase focused on determining why and when a given component would be procured outside the firm, in a market transaction, rather than supplied internally from a wholly-owned subsidiary, through hierarchical coordination governed by transfer pricing. (See for example, Coase 1937; Klein, Crawford, and Alchian, 1978). Oliver Williamson’s (1975, 1985) answer to the question primarily involved “asset specificity” -- when investments in firm-specific assets were required, economic logic pointed towards maintaining vertical integration, i.e. internal manufacturing, of those components. Under conditions of low asset-specificity -- commodity parts of one sort or another -transactions in spot markets or short-term contracts based on low-bid competition were superior for obtaining the best price for a given level of quality. The shorthand for this decision-process was “make vs. buy”, highlighting the differentiated advantages of hierarchy and market as methods of coordinating economic activity. Roughly fifty years ago, another supply chain phenomenon re-appeared on the global automotive scene. Certain automakers procured parts externally, rather than through vertical integration, but not in spot markets or through short-term contracting. Instead, they relied upon a small number of supplier firms with whom they had long-term business relationships (often including an equity stake) – “relational” contracts governed by understandings about sharing 2

both pain and gain, and large amounts of asset-specific knowledge on both sides. This approach allowed the close coordination on design and manufacturing tasks usually associated with vertical integration, while maintaining the potential for price pressure and supplier competition associated with market transactions.1 This development shifted attention from “make vs. buy” to the different issue of “how to buy,” which deals centrally with the terms of the relationships among the parties. Hirschman (1970) famously distinguished between economic relationships managed by the constant threat of “exit” from those managed by “voice,” an exchange of views within an ongoing relationship.2 Helper and Sako (Helper, 1991; Sako, 1992; Helper and Sako, 1995) apply these terms to the dominant modes of supply chain management associated with the U.S., whose “exit” approach involves arms-length relationships, selection based on low bid, frequent switching among suppliers, reliance on contracts for governance, as opposed to Japan, whose “voice” approach involves long-term relationships, selection based on supplier capabilities, frequent collaboration within stable supplier partnerships, reliance on normative understandings for governance. In the language of the Adler-Heckscher chapter (chapter 2), the buy-exit option relies on market, and the buy-voice option relies on community; the make option could rely either on hierarchy (as in a traditional vertically-integrated organization), or on community (in a networkbased firm). The application of “exit” and “voice” to supplier relations captures the orientation of the parties towards their relationship (as opposed to just a single transaction) and frames the issues of information-sharing that crucially affect the relationship beyond transaction-specific costs. “Exit” is characterized by the creation and exploitation of information asymmetries by both parties, even when the relationship endures over long periods of time. “Voice” requires shared


The transformation in supplier relations Over the last half century four trends -. As we discuss below. the hard-nosed “exit” approach of U. we will argue. the industry is converging from all sides on a form of pragmatic collaboration.norms of reciprocity that balance the willingness of the customer to undertake investments in the supplier’s capabilities against the supplier’s responsibilities to invest in new technology and capacity. There has been a wide range of responses to these pressures.each characterized by a different rate of change -. “voice” distinction is no longer as clear as it was just twenty years ago. Thus. global competition brought Japanese vehicles (beginning in the 1960s) and Japanese manufacturing facilities (beginning in the 1980s) to the U. though more open and formalized than the traditional Japanese system. but this approach is marked by internal contradictions which. depending on such conditions as firm strategy and market structure. On the other. In the U.S. both exit and voice have been profit-maximizing strategies in the past. requiring more formalization and costjustification of the relationships. there are frequent attempts to achieve the necessary levels of collaboration without trust. involving substantial levels of trust. firms has faced pressure for increased collaboration to achieve the increased levels of quality demanded in the market. Yet the “exit” vs. the closed keiretsu system of suppliers characteristic of Japanese industry has been considerably opened to market pressures. On the “voice” side.have had a profound effect on relations between automakers and suppliers. First.S. The need to compete with Japanese automakers on quality and the gradual diffusion of lean production created incentives 4 .S. often mixed and contradictory. we believe. make it unlikely that it can stabilize as a lasting model.

requiring a great deal of ongoing communication among the designers of different parts. automakers to increase product quality. Given a greater availability of parts 5 . The third factor is less a trend than a continuing reality amid other dramatic changes. Since the establishing of a dominant design in the 1930s. perceiving increased competitive advantage from focusing on core competencies.for U. automakers and mega-suppliers alike began determined efforts to move towards a more modular approach. By outsourcing both manufacturing and design. while also reducing labor costs. context.S. Second. given a growing gap between wages and benefits at the automakers’ in-house parts divisions (which were unionized) and at non-union independent suppliers. The subsequent decision that many design tasks should also be outsourced resulted in a much more rapid pace of deverticalization (as measured by the percent of value-added outsourced by the automakers) and the creation of a new breed of megasupplier. Fourth. Starting in the mid-1990s. vehicle architecture has proved stubbornly resistant to these efforts and retains a high level of integrality. this has given automakers greater leverage over most suppliers in price negotiations. Outsourcing of manufacturing was the initial focus. In the U. automakers could rely on specialized supplier expertise. and achieving this increased quality required a more closely coordinated relationship with suppliers vis-à-vis design. thus defying expectations of more independence for suppliers during the design process. and parts production. the product architecture of the automobile has been primarily integral. global overcapacity in automotive assembly and the parts sector has increased price pressure on both automakers and suppliers. rather than maintaining that expertise in-house. starting in the 1970s. However. there was a trend away from vertical integration (deverticalization) in the US.S. sub-assembly. Automakers also turned away from vertical integration for strategic reasons.

where suppliers are selected and contracts are written.task-level collaboration without the underlying relational and value base that cements longer-term relationships -. a “China price. contrary to customary expectations for collaboration. and 2) Does collaboration necessarily require trust? These questions bear on the central question of this volume. collaborate on certain core engineering. and has contributed to organizational inertia with respect to moving towards a more collaborative mode of exchange. the level of collaboration between automakers and suppliers has increased. Other authors here. Observing the confluence of these developments over the past ten years. in response to these developments. and product design tasks while at the governance level. We take this position despite other developments during this same period that have appeared to push the auto industry in the direction of reduced collaboration. especially in the US. there is an adversarial relationship and lack of trust.” This reinforces the apparent value of U. automakers’ long-standing purchasing routines built around exit. these supply chain relationships do not always involve high levels of trust. argue that effective collaboration requires trust 6 . some firms.S. manufacturing. That a viable model. we conclude that the underlying forces affecting supply chains have in fact made collaborative relations more important rather than less. automakers have more options for exit and hence a more credible threat in demanding that their existing suppliers meet. for example. As a consequence of these four trends. This phenomenon poses two questions: 1) Can organizations with an exit-oriented tradition create effective forms of collaboration?. particularly Adler and Heckscher in chapter 2. We will explore to what extent this model of “collaboration without trust”3 -.from newly-sophisticated suppliers in less-developed countries. Yet we have been surprised to see that.

that thesis is called into question.g. consumers did not have the ability to buy higher-quality cars. 1992). Therefore. since suppliers usually did not design their own products. when the Japanese entered the U. automakers was maintaining a credible threat of exit. one bracket rather than an entire headrest assembly). However. This short time horizon and extreme division of labor resulted in inefficiency and poor quality. partly by outsourcing only simple manufacturing tasks (more complicated tasks like sub-assembly and design stayed in house) and partly by standardized specifications and bidding procedures. 1991.S. To the extent that the benefits of collaboration.S. often with multiple suppliers per part. Despite their inefficiency. consumers did gain 7 . quality.However. it was difficult to optimize across components. automakers had similar practices. and lean production To an ever-increasing extent in the post-war era. with production of simple parts done by outside suppliers under short-term (1-year) contracts. automakers created a large supply of potential suppliers. exit relationships therefore maximized automaker profits by making it easy to switch suppliers (Helper. Helper and Levine. first with imports in the late 60s and early 70s and soon with local manufacturing plants in the early 1980s. they could not optimize them for their own production processes.S. The key consideration for U. can be achieved without trust. market. industry operated in exit mode.S. such as flexible resource allocation and response to change.and institutions of community to support it. Global competition.4 Since each supplier produced only a small component (e.. to prevent “hold up” by suppliers taking advantage of assetspecific knowledge developed over time. For example. since each of the Big 3 U. the U.

promoted the idea that in-house parts production had become inefficient during years of Big Three quasi-monopoly of the U. Instead. Deverticalization and the emergence of “mega-suppliers” In the 1970’s. they needed more capable suppliers that combined a variety of skills. Why did vertical integration decline? Three factors seem to be responsible: 1) A new breed of purchasing executive. To do this.more outsourcing to reduce reliance on in-house divisions and intensified bidding procedures to force greater price competition among suppliers. The re-integration of design and production could have occurred by taking more production back in-house. The Japanese quality advantage was based on thoroughgoing adoption of “lean production” practices governed by voice supplier relationships. the Big Three had to improve quality by re-uniting design and production and increasing the size of subassemblies. the opposite occurred—design gradually moved to suppliers. instead of the “bend and send” “shoot and ship” firms with few design or management skills.access to more reliable cars. component design was largely done by the automakers in-house. purchasing moved towards a more aggressive use of market mechanisms -. although production was done by a mix of vertically-integrated divisions and financially-independent firms. and the Big 3 spun off their parts divisions.5 To compete with the Japanese.S. exemplified by Ignacio (Inaki) Lopez at General Motors. 2) The wage gap between non-union independent suppliers and unionized in-house supply divisions grew during the 1970s. To remedy this. automotive market and that long-standing relationships with outside suppliers were “cozy” and riddled with waste. capable only of doing one narrow production process. as suppliers grew more bold in their union-avoidance 8 .

Magna. aggregations (or “chunks”) of components. The interface between design and production can be organized in three ways (Clark and Fujimoto. TRW Automotive.. Denso. By the mid 1980s. and assuming a larger role in the management of second and third-tier suppliers.e. 3) Managers perceived an increase in returns to specialization. which were able to expand their nonunion operations rapidly. The UAW was able to maintain its strength at the Big Three by threatening to shut down assembly plants. respectively. 1991).strategies and inflation eroded the purchasing power of non-union workers. i.5).. handling more complex manufacturing and logistics tasks. This increased reliance on outsiders was bolstered by new management theories (e. Prahalad and Hamel. the wage gap between a worker in a Big 3 component plant and an independent plant ranged from 2:1 to 3:1 (Herzenberg.g.6 The pace of deverticalization increased dramatically when first GM and then Ford spun off their captive parts divisions. More mega-suppliers were soon on the way. but did not have such leverage over independent parts suppliers.7 More subcontracting does not necessarily mean more collaboration. they chose to rely on the expertise of outside suppliers from those industries rather than expanding their in-house knowledge. formed from horizontal merger and acquisition activity in order to compete for larger subassemblies. Eaton. Lear. Federal Mogul) now growing larger and taking over critical design and engineering tasks. 1991): supplier proprietary (supplier designs and manufactures the part and sells it through a catalog). These were mostly existing automotive suppliers (Johnson Controls Inc. 9 . As automakers began to take advantage of innovations in electronics and plastics in the 1980s. Dana. Delphi and Visteon immediately became the first “mega-suppliers” (also known as Tier 0. 1990) which argued that firms sticking to narrow core competencies performed better. creating Delphi (1998) and Visteon (1999).

S. and Europe. and Japanese companies in how subcontracts are organized. case. In that context. in contrast. with 62% of all procurement cost handled in “black box” mode in Japan vs.OEM (Original Equipment Manufacturer) detail-controlled (all design specifications are predetermined by the OEM. 30% in Japan. and Japan. vs. a low level of vertical integration was synonymous with a high level of “black box”.”8 In data from the late 1980s. and 81% of procurement cost handled in OEM detail-controlled in the U. respectively.S. “black box” subcontracting involves the greatest collaboration of the sort that we term (below) “pragmatic. etc.S. From this perspective.S. The decision by OEMs to pursue lower labor costs and to focus on core knowledge and capabilities initiated this move. and the supplier provides the rest of the design). with 54% detail-controlled parts and 39% “black box”. and the resulting 10 . reveals that even as vertical integration dropped from its peak in the mid-1950s. European firms were in between. As these percentages suggest.S. <<Tables 1 and 2 about here>> The recent wave of deverticalization increases the reliance on “black box” subcontracting as a replacement for the OEM detail-oriented approach. or “black box” (the OEM provides performance requirements and basic parameters of size.S. data are chosen to reflect the impact of the Delphi and Visteon spin-offs. the U. the Japanese data are more recent (and more precise). Clark and Fujimoto find marked differences between U. the detailoriented approach has been quite consistent with the exit mode of exchange.S. weight. the supplier has no design role and only manufactures the part). The U. 16% in the U. Tables 1 and 2 reveal levels of vertical integration over time for the U. collaborative subcontracting. the resulting subcontracting was not often collaborative. the Japanese subcontracting system never achieved the same level of vertical integration in the auto industry as the U.

2002). is efficient in part because it requires little collaboration.9 Thus modularity.creation of mega-suppliers with enhanced design capabilities accelerated the trend. Any PC manufacturer can hypothetically use any hard drive from any supplier. as long as the predefined function is fulfilled and the standardized interface is present. Beginning in the late 1990s (and influenced by the success of companies like Dell that take advantage of modular product architecture to “build to order”).” while an integral architecture requires intensive coordination throughout the design process. Table 3 documents the growth of mega-suppliers beginning in the mid-1990s and Table 4 lists the top 10 mega-suppliers as of 2003. automakers began to think 11 . automakers and suppliers began anticipating the rise of “modularity” as the new basis for designing automotive products. Put differently. whereas “integral” product architecture is based on interdependence both within and across elements. without extensive coordination with innovations in other modules (Langlois. <<Tables 3 and 4 about here>> “Modularity” and the predicted decline of collaboration One rationale for the creation of mega-suppliers was a prediction about the future of automotive product architecture. Influenced by the example of information technology. established by the architecture’s pre-determined “design rules. modules can be developed independently from one another and connected via standardized interfaces. “Modules” as a basis for product architecture are defined as elements that are “interdependent within. Because a module provides one pre-defined function and its interface compatibility with other modules is assured through standardization. according to its proponents. and independent across”. Modularity may also reduce asset specificity. innovation within modules can proceed independently.

But the move towards modularity has been much less decisive than its advocates predicted a decade ago. unlike the production of computers. problems that can result in noise. Related to this is the need to avoid problems that result from the interaction of the parts. vibration. Similarly. Automotive product architecture. A car is vastly more complex than a PC. has proved resistant to moving away from integrality. space is at a premium. 2004). in today’s dominant design. mega-suppliers brought proposals for module designs to their customers. but usually not even across models designed by the same automaker). This violates formal definition of “modularity” in multiple ways: more than one function is mapped to the “chunk”.of how they could divide up the vehicle into discrete modules. so components are designed to conform tightly to model-specific physical constraints. What is called a “module” could be more accurately described as a chunk of physically proximate components that could be sub-assembled independently from the rest of the vehicle. and there is no standardized interface allowing interchangeable connectivity of modules. and then installed on the final assembly line in a single step. laptop 12 . it means something quite different than in the information technology world (Sako. and it is highly dependent for marketing on a distinctive visual identity. and harshness of ride. tested for functionality after subassembly. some of which could be fully outsourced to mega-suppliers. there is no standard definition of the functions performed by a module (certainly not within the industry.10 Why has it been so hard to modularize the car? First. At the same time. it must use space much more tightly. For all these reasons the plugging together of standardized pieces works less well than in the computer industry. Though the term “module” is often used today in the auto industry.

although predicted to be lower due to economies of scale from standardization -. Furthermore. differences in design philosophy meant widely different numbers of defined modules and no agreement on modular boundaries.have often proven to be higher than the collection of components. Achieving systemic integrity for these functions requires precise coordination across components to meet requirements for a vehicle of a particular size. center of gravity. Even aside from these industry-specific reasons. even for products whose architecture is easily decomposable into 13 . weight. Third. individually installed. and supplier capabilities are in some cases not adequate to the design responsibilities given them.computers are more integral than desktop PCs because of the need to utilize scarce space efficiently. even within a single automaker. Some reasons are: OEMs preserve a “shadow engineering” presence to monitor supplier engineers. many important functions or subsystems are geographically distributed around the vehicle. the “look and feel” of a given vehicle is important to brand image and the emotional connection to the customer. that they were meant to replace. OEMs are reluctant to allow a first-tier supplier to choose secondand third-tier suppliers due to the volume-based contracts the automakers have negotiated directly with these suppliers. there is now a greater realism about the inherent limits of modularity. etc. and designers fear loss of distinctiveness from moving to the use of standardized modules. thus constraining product innovation and cost reduction. braking. Second. when various companies began trying to divide up the vehicle into a set of standard modules. such as safety. electrical. even as current purchasing routines prevent suppliers from including investment costs for those capabilities in their piece price. steering. the costs of automobile modules -. Finally. Engineering costs for modules can be higher.

modules. first-tier suppliers. of the mega-suppliers. increased. these requirements have. In fact. Although the shift towards collaboration was noteworthy during the 1990s. the importation of the concept of “module” has had value for automakers and suppliers by causing them to think about larger chunks of components as the relevant unit for design and sourcing decisions. given ongoing changes in underlying technologies and consumer demands for functionality. Despite this litany of problems. legacies of exit remained in incentive systems and compartmentalized 14 .13 Legacies of exit and global overcapacity. through mergers and acquisitions. and (by extension) lower-tier suppliers as well.11 Such agreement may be impossible to achieve (or to sustain over time) at any meaningful level of aggregation. one thing that definitely did not occur was any reduction in task interdependence and coordination requirements between automakers. It has also provided the rationale for changes that ultimately. have more to do with deverticalization than with any change in product architecture -.12 Whatever can be said about the changes wrought by the arrival of “modules” in the automotive industry. due to the combination of the shift in design responsibility for components associated with deverticalization and the move towards designing and manufacturing bigger “chunks” as discrete units. or simply a lack of willingness in a competitive context to sacrifice proprietary or brand-influential architectural features in order to achieve the gains from standardization. we would argue. if anything. because of the requirements of achieving advance agreement about module boundaries and a standardized interface.including the very formation. The persistent integrality of automotive product architecture has made necessary continued intensive collaboration between assemblers and suppliers. it was far from complete.

but it does temper it. but rather was sold cheaply at auction to firms that continued to produce. Despite wage cuts. The temptation to return to exit was intensified by over-capacity. 1993). this excess supply does not obliterate the tendency toward collaboration that results from greater global competition in final product markets. In the meantime. (Sturgeon. automakers will find themselves with a variety of suppliers to choose from. Purchasing agents continued to be rewarded for their ability to cut suppliers’ piece prices. and not so much for their ability to ensure the on-time delivery of high-quality parts. it may take another decade. but given the long life of both workers and equipment. Much of the equipment used by plants that did go out of business was not scrapped. 2002). making exit on some occasions tempting even for practitioners of voice. improvements in transportation and communication combined with low wages in less-developed countries made it attractive for multinational firms to build new plants in nations such as Mexico.organization. particularly those with high-school educations. Japanese automakers frequently brought their suppliers with them to the US. often finding it easier to bring their own suppliers than to train US firms in the techniques of lean production. Eventually. This increase in new capacity was not matched by shutdowns of old capacity. As we discuss below. Meanwhile. China. supply and demand will come more into balance. particularly among parts-makers. in the 1980s over 300 Japanese auto suppliers came to the US (Kenney and Florida. many workers. found that staying in the industry was their best alternative. 15 . and India.

Rather suppliers are competitively evaluated and cessation of business is not uncommon. Carried over 16 . Given the higher costs (both perceived and real) of vertically-integrated suppliers and the persistent integrality of vehicle product architecture. although less frequently and speedily than under “exit. conversely. Thus firms with an “exit” legacy find themselves needing to develop collaborative capacity in response to deverticalization.”.” The fundamental orientation of the parties is long-term and relational. but nor is it based entirely on assessment of capabilities within a closed group of suppliers. a wider and more open range of relationships than the traditional keiretsu. and unfamiliar demands from new customers -. customers are open to establishing relationships with new suppliers. Selecting a supplier does not involve bidding. we see an increasing degree of collaboration in automotive supply chains over the past 10-15 years. as in exit. the typically Japanese “voice” pattern has been confronted with greater competition from new entries into the once-closed group of suppliers. as in voice. The typically American “exit” pattern has been pushed in the direction of longer-term relationships. This form of collaboration no longer fits cleanly into the “exit” and “voice” strategies that characterized the earlier phase. automakers found this pursuit of quality took a form that required more coordination and collaboration on design and production with their financially independent first-tier suppliers. However. Table 5 elaborates the hybrid collaborative mode of exchange in relation to “exit” and “voice. greater demands from the global expansion of existing customers. This assessment continues even after the relationship is established. and firms with a “voice” legacy must be more prepared to face competitive pressure. as with short.The shift towards collaboration To summarize: due to the four trends described above.

which was very unusual in the past. These more general (less customerspecific) techniques make it easier for customers and suppliers to collaborate effectively even when the relationship is relatively new.g. the supplier’s design role is much larger. mirrored by an increased share for Supplier B. it also relies upon extensive use of formalized process management routines affecting problem-solving in manufacturing and design rather than simply upon tacit understandings based on long-term relationships. the hybrid mode finds variation. Under the hybrid mode. While equity stakes were common under voice and not under exit. while exit codifies procedures and voice relies on tacit understandings. Poor performance by Supplier A may not cause exit but a reduction in the share of the customer’s business. but otherwise this is not common. While the hybrid mode relies more heavily upon dialogue than formal contracts for governance. Automakers frequently take equity stakes in suppliers who are expert in technologies that are assuming an increased importance in vehicle design (e. yet the supplier role in developing new designs was limited. composite materials). <<Table 5 about here>> While design under exit is simplified in order to generate a larger pool of potential suppliers. to the extent that customers express high concern about the level of a supplier’s design capabilities -. the hybrid mode relies on process management routines that make procedures explicit. design under voice was typically organized so that supplier engineers could be “resident” at the customer’s design facility. electronics. Similarly.from voice is the manner in which performance problems are handled. 17 .

a relatively small departure from the history of “black box” subcontracting within a voice mode of exchange -. even in the midst of established contracts. We explore the situation in Japan in greater detail as one of three case studies below. OEMs have been caught taking those incipient designs and sending them to a another supplier to get a competitive bid. Given the tradition of working closely with suppliers as collaborative partners. using online “reverse auctions” in which some of the bids pushing prices to record low levels could not be verified as coming from credible suppliers. there remains substantial variation. OEMs began telling suppliers they would now be responsible for those costs. 18 . Here we describe the range of responses and below we provide interpretations of the U. The American Big Three’s interpretation of collaboration has often been to impose increasing requirements on suppliers. Within this general move towards increased collaboration.although at some companies. and. For example. patterns. Much of this behavior is more extreme even than past norms of the “exit” mode. Other extreme examples include OEMs demanding immediate 5% price cuts. however. confronting suppliers with a “China price” and demanding they meet it or lose the business. the hybrid mode takes the form of “collaboration with trust”. regardless of contractual terms already established.Types of collaboration: with and without trust14 It is our contention that convergence towards this hybrid mode of collaboration has characterized automotive supply chains over the past ten years and that this trend will continue. the implications for the traditional keiretsu structure of the supply chain are quite significant. The greatest continuity with previous practice can be seen in Japan.S. where the most variation is found. with suppliers taking a larger role in providing component designs as part of their bid. Breaking from a long tradition of paying for tooling as part of a supply contract. most recently.

S. transaction-based behaviors have at times provoked strong public reactions -. While supplier relations at the level of specific design 19 . Honda. You know that the suppliers raked over the coals and used as a whipping boy to explain the Big Three’s cost problem are the same suppliers investing. But it is also very clear that our futures are inextricably tied and neither can afford the other to fail. Michigan in August 2002. 2002 Responding to this cry from the heart a few days later.15 This excerpt captures the strong emotions stirred up by these developments: There is little chance that beating the hell out of the supplier base and breaking contracts… is going to get to the root cause of your problem. President/CEO of Metaldyne. GM’s CEO Rick Wagoner had a ready answer: “Stop whining!”16 Leuliette’s speech points out a powerful underlying dynamic: When given the choice. and Nissan.Such hard-nosed. Governance-Level Divergence The descriptions above suggest that our earlier statement about a convergent trend towards collaboration must be qualified. and earning a good return with the vehicle producers that have the growing market share. August 12. There is a discontinuity here. so many of them have adapted as best they can to operating under these new conditions. suppliers have been learning to prefer working with the Japanese transplant manufacturers who operated in “voice” mode when establishing supplier contracts -even as their Big Three customers came to adopt more and more of the design and manufacturing practices long in place at companies such as Toyota. Tim Leuliette. suppliers will readily choose collaboration with trust over collaboration without trust. One noteworthy speech was made by the CEO of a mediumsized supplier at the annual industry briefing meetings in Traverse City. Task-Level the point where any suggestion that collaborative activity might exist between these adversaries can seem absurd. but steadily. building partnerships. Gradually. Big Three. Yet few suppliers can afford to turn their back on Big Three business. U.

Indeed. internal lack of coordination at the OEM. at the firm level. At the governance level. Some firms are choosing a greater reliance on market mechanisms to keep price pressure on their suppliers. who is responsible for warranty and product liability costs. it is both the case that managers in charge of sourcing seek to maintain a diversity of in-house capacities and subcontracting relations AND that different strategic sourcing practices compete with one another for dominance. At times these pressures can lead to contradictory behavior in different arenas.and manufacturing tasks may be responding similarly to the forces analyzed above. the responsibility for investments in tooling and other capabilities. both seem to distribute their strategies to accommodate as broad an array of (even contradictory) sourcing strategies as possible. past history of “exit” or “voice” doesn’t necessarily determine which path is chosen. Within OEMs. the differential importance of one criterion (price vs.17 Herrigel (2004) summarizes this perspective: Are the new relations cooperative and collaborative. some thoughtful observers suggest that there is no single emerging pattern but rather a portfolio of viable OEM-supplier relationships whose diversity is motivated by various things: risk-hedging. and the way that disputes are handled -can all vary widely. the policies that govern the transaction -. This is because neither the actors in OEMs nor the actors in component producing firms make such drastic choices. the negotiations over price. design) for that particular part. or are they still essentially about cost and price? … [It] is important to avoid the urge to choose between these hard alternatives. Interestingly. At one limit. In reaction to this de facto multiplicity of OEM sourcing strategies. bidding for work). suppliers’ protective moves. Reflecting on these developments. these tasks are carried out within differing transactional contexts. from the supplier’s perspective. we find instances where collaborative activity occurs at the task level.the process of awarding work to a supplier (or. OEMs subject suppliers to severe versions of “exit” behavior. while at the governance level. quality vs. while others are choosing to pursue a small number of longerterm relationships within which various issues are resolved. 20 .

For example. maintaining a credible threat of exit is facilitated by an ability to make ‘apples-toapples’ price comparisons. This in turn requires designs to be well-documented (so suppliers know what they are bidding on). the choice of strategy affects not just purchasing but the entire corporation.component producers are developing a broad range of strategies that take advantage of the (sometimes quite unpredictable) variety of OEM sourcing practices. voice benefits from agents with engineering backgrounds and complex compensation schemes that balance multiple objectives. First. and suggests that purchasing will act as a gate-keeper.18 Indeed. exit and voice require different criteria for choosing and compensating purchasing agents: exit strategies benefit from financially-oriented purchasing agents who are compensated based on their ability to keep prices low. any given pattern of relationships leads to differences all the way through the organizational system. limiting suppliers’ access to the engineering department. Second. Collaborative “cost-down” activities can achieve both. inherent in our view of “voice” (and of the hybrid mode) is the idea that the parties can actively circumvent any necessary tradeoff between collaboration and efficiency. under voice much 21 . where they might succeed in obtaining design changes that would limit purchasing’s ability to compare prices. consider that economic relationships are based on mutual expectations resulting from past experience: if a pattern of low-trust expectations is established it cannot easily be shifted to high trust for a new transaction. In contrast. Like cars. For example. To explore the problems of “mix and match” strategies. Furthermore. 45-46) We have two objections to this view. patterns of relationships are integral. (p. we are skeptical that supplier relationships can be constructed or maintained by “mixing and matching” contradictory strategies.

partly because the process of dialogue and debate allows “learning by monitoring” (see chapter \Sabel\). Those individuals engaged in interdependent design and manufacturing tasks can assure themselves that even as they are advancing completion of the task. and collaboration with engineering is considered a good thing. while there can be some variation in how individual suppliers are treated. Thus. the interdependent design and manufacturing practices linking automakers and their suppliers at the task level are pragmatic in that they confront the uncertain and changing nature of knowledge with a continuing willingness to question current routines and past choices and to explore alternatives.19 The distinction we draw between collaboration “with trust” and “without trust” reflects the consequence of superimposing the new reality of iterated co-design upon the legacies of “voice” and “exit” modes of exchange that continue to exert inertial influence on purchasing routines and governance arrangements.knowledge can remain tacit (not documented) since relationships change infrequently. they are able to 22 . we question whether the pattern of task-level convergence and governance-level divergence can be explained by deliberate efforts by firms to sustain a portfolio of different purchasing strategies. Emerging alternatives: pragmatic collaboration. we see two emerging alteratives that account for the range of observations above: a convergent trend towards “pragmatic collaboration” at the task level that can be found within divergent approaches at the governance level. <<Figure 1 about here>> In this view. Collaboration emerges naturally during this exploration. with and without governance trust Rather than a panoply of strategies. Figure 1 displays the underlying logic generating this pattern.

since they too recognize the gains in the face of technological change and unpredictable market demand. The problem for such firms is the absence of trust at the governance level. the very process of pragmatic collaboration generates the levels of task-level trust necessary for the process of iterated co-design -. We return to these issues below. 2) the coping strategy of Stoneridge. in our view. non-opportunistic fashion. and the difficulty of developing a form of pragmatic collaboration at that level capable of generating trust over time. Case examples We turn now to consider three individual case examples: 1) the continuing phenomenon of “collaboration with trust” in Japan. although we see a number of reasons why it will not disappear quickly. as practiced by firms like Toyota and Honda and the recently different approach of Nissan. a small second-tier supplier faced with heavy downward pressure on profit margins from its Big 3 customers. We do not expect that this internal contradiction can be viable over the long term.and for persistence of task-level relationships. Thus.ascertain whether the other party is behaving in trustworthy. and 3) the effort by mega-supplier Delphi (formerly part of GM and hence with a strong legacy of “exit”) to change its purchasing and supply chain management practices in order to move towards the “collaboration with trust” model. 23 . The pragmatic evolution of collaboration opens the door. we have argued. for participation in such collaborations by firms with an “exit” history. Starting from the premise that the pace and extent of deverticalization and lean production are requiring more collaborative interdependence between automakers and suppliers.

the first case considers how Japanese firms have updated their former “voice” approach to what we are calling “collaboration with trust.” The second and third cases consider the U.S., where OEMS and suppliers alike have a legacy of “exit” relations with low trust. Here we find contradictions, with an increase in trust-destroying behaviors by OEMs on the one hand -- as in the second case of Stoneridge, a small U.S. electronics supplier -- while on the other hand some mega-suppliers attempt to move towards “collaboration with trust” – as in the third case of Delphi. The cases will set the stage for a final discussion of what kind of supplier relationships we are most likely to see in the future, both in the U.S. context and beyond. “Collaboration with Trust” as Work in Progress -- Supply Chain Management at Toyota, Honda, and Nissan Toyota and Honda are rightly regarded as the leading practitioners of “collaboration with trust” in their relationship with suppliers, although the earlier characterization of the “voice” mode of exchange was often applied to all Japanese firms. We include Nissan in this assessment because of its much-heralded recovery from near-bankruptcy after an alliance with Renault and the importation of a Western CEO (Carlos Ghosn) who set out energetically to dismantle key aspects of the old supplier system. Taking a current snapshot of these firms is a good way to consider “collaboration with trust” as a work in progress. We will argue that current manifestations of “collaboration with trust” are not simply the previous “voice” mode of exchange, whose Japanese form was highly intertwined with the industrial structure of interlinked firms known as keiretsu. Rather, they are a modern adaptation reflecting the forces and pressures affecting the entire industry, as described above, as well as the very substantial foreign direct investment by leading Japanese automakers in establishing a manufacturing footprint and building local supply chains all over the world. Globalization has 24

required collaborative processes that are considerably more accessible to out-of-network suppliers and contain more explicit articulation of norms and values as well as specific guidance in how to approach task interdependencies. Japan’s prolonged recession in the 1990s -- what some observers call the “lost decade” -- put tremendous strains on both first-tier and lower-tier suppliers in Japan. First-tier suppliers were often asked to join their OEM customers in expansion efforts overseas, opening new plants first in the U.S. and then in Europe. At the same time, these firms were investing heavily in Southeast Asia in order to reduce their production costs, but this severely strained the lifetime employment commitments at their domestic plants and sparked social criticism of their role in the “hollowing out” of the Japanese economy. Smaller second- and third-tier suppliers faced a starker challenge. Japanese OEMs typically pledged to invest heavily in supply base development when entering a new country and they (and their first-tier partners) could easily replace the small Japanese supplier by sourcing their parts from domestic suppliers near their new overseas manufacturing plants. While these small suppliers were not cut off completely, many struggled to stay in business. The general economic conditions are somewhat improved now, but the fundamental challenge to small Japanese suppliers to justify their continued place in the supply chain remains. During this same period, the past pattern of strong vertical control of design activities and production coordination by the OEM has begun to change. This raises the same fundamental dilemma: OEMs want suppliers to take a larger role in design and in managing lower tiers of the supply chain, but they also want to maintain their knowledge and power; they recognize that they must “know more than they make.”20 Suppliers must therefore make more investment (in capacity and knowledge), which is potentially redundant and not paid for by OEMs.


Japanese OEMs have also put suppliers under growing pressure for cost reduction. Nissan took the lead, with Carlos Ghosn setting forth a drastic cost-cutting plan that aimed to shock the old keiretsu system into changing past patterns. Toyota, determined not to let Nissan gain any cost structure advantage, followed soon with its CCC 21 program (Construction of Cost Competitiveness for the 21st Century) that aimed for 30% reductions in parts prices -- targets much higher than the usual incremental “cost-down” targets.21 Akira Takeishi (2003) and his colleagues argue that Japanese OEM-supplier relations must move into a new fourth phase, after three previous phases that emphasized cost, quality, and engineering input for individual components respectively. The newest phase emphasizes horizontal coordination with suppliers of other components for improved design, greater system integrity, and more innovation at both component and system levels, as well as exploration of more modular design principles. The Japanese word kyogyo is used to describe this phase; it means “collaborative division of labor.” OEMs have shown a willingness to turn more design responsibilities over to suppliers who can manage these horizontal collaborations most successfully. Certain examples of kyogyo are well-publicized in Japan, e.g. a instrument panel console for a new Lexus model that resulted from the self-initiated collaboration of Sumitomo Denko (an electronics firm) and Toyota Gosei (a plastics firm) and that achieved sizable improvements in terms of lighter weight, lower parts count, and lower overall cost. Initial expectations were that kyogyo might be the first step towards modularization, but this particular example suggests otherwise, for the Lexus console is completely idiosyncratic to a single Lexus model. It is possible that these suppliers will attempt a more standardized design that they can offer to multiple customers, in order to spread their fixed costs over more units.


often through direct supplier development activities that began in the 1990s and continue to the present day.22 Overall. supplier development in the U. these are considered to be a better method of supplier development in terms of sustainability. they emphasize short-term fixes that produce immediately visible savings. and shorten lead times. Hajime Ohba.S. increase logistical accuracy. They do not want us to visit their plants and they are unwilling to share the information we require. where the integrality of product architecture is accepted as fundamental to how product development is organized. 673) More recently. called Jishuken in Japan. Within overseas operations.S. Thus “collaboration with trust” within Japanese supply chain structures is likely to take on more of this “first horizontal. suppliers do not understand our way of doing business. we also can’t help them improve. since the direct supplier assistance by the OEM often produces too much dependency. locates the reason in the overwhelmingly financial orientation of many U.” (p. it was difficult for Toyota and Honda to find suppliers who would agree to participate in their collaborative approach. then vertical” coordination flavor in the future.Within the Japanese context. head of Toyota’s Supplier Support Center in Kentucky. Toyota and Honda continue to strengthen their local supply chains in terms of systematic production capabilities to reduce cost. This makes it very difficult for us to work with them effectively. has been a slower and more difficult process than originally envisioned by the Japanese OEMs. typically through inventory 27 . Dyer and Singh (1998) quotes Toyota’s VP for purchasing Koichiro Noguchi from Dyer’s interview with him in 1992: Many U.S. At first. suppliers. we think it more likely that these horizontal collaborations will continue to result in highly integral designs that still require high vertical design interdependence with OEMs. improve quality. Slower to develop have been supplierto-supplier working groups.

24 Renault provided that investment in March 1999. and don’t persist with more fundamental changes in how they approach manufacturing. U. As these automakers localize more design activity to their R&D facilities in the U. such as Toyota. Nissan held equity stakes in keiretsu companies totaling over $4 billion. with evidence that overall costs are lower and the amount of innovation contributed is higher.reduction. and Europe.8% stake in Nissan and thus freeing up $5. At the time of the alliance. Retained earnings were typically invested in purchasing shares of these affiliated companies and supplier CEOs were routinely appointed from among the ranks of senior Nissan executives approaching retirement. suppliers are increasingly vocal about their preference for working with Toyota and Honda rather than the Big Three. more design involvement by local suppliers will result.23 As noted above. became the new CEO. with both Moody’s and Standard and Poor’s announcing plans to reduce its credit rating to “junk” status unless it received outside investment. cultivating extensive financial and personnel interconnections with its suppliers. Nissan was in deep financial crisis. assuming a 36. One target of Ghosn’s cost-cutting was Nissan’s keiretsu ties with suppliers. well-known for his success in cost-cutting and turning around troubled operations. Nissan had followed a traditional keiretsu approach. Yet these cross-sharing holdings and personal relationships did not yield the cost advantages achieved by Nissan’s competitors. 28 .4 billion in capital and retaining “investment grade” bond status. Nissan’s situation demonstrates some of the gains and losses that occur when a company with a “voice” tradition resorts to “exit” behavior.S.25 Carlos Ghosn from Renault.S. In early 1999. already. By 1999. Nissan’s purchasing costs were estimated to be 20-25% higher than Renault’s costs.

called Nissan 180 to reflect an increase of sales by 1 million vehicles. One reason was platform consolidation. To be achieved by April 2005. divesting itself of all other businesses and hence moving from a first-tier to a second-tier position in Nissan’s supply chain. as part of the Nissan Revival Plan (NRP). but with high expectations for ongoing cost reductions. Nissan 180 called for a further 15% reduction in purchasing costs while also 29 . The number of suppliers shrank by 40% overall. Supplier companies often made dramatic changes in response to Nissan’s new policy. With this introduction of “exit” mechanisms. which also included internal plant closings. four plants produced vehicles based on 15 platforms. prior to NRP. From most accounts. and zero debt -. this change of policy was implemented remarkably quickly and smoothly.was even more ambitious. one long-time supplier of brake systems chose to focus on just one group of components. These teams. had primary responsibility for developing recommendations for how to achieve the goals of the turnaround. for example. Nissan’s purchasing costs declined by 20% by March 2002.Ghosn and the Nissan board reached an early decision to end its equity participation in keiretsu supplier companies and to accept competitive bids from outside suppliers. Keiretsu suppliers were still encouraged to compete for Nissan’s business. he won internal support for these policy changes by creating a set of cross-functional teams that reported directly to the Executive Committee and were given access to all company information. Nissan had seven plants producing vehicles based on 24 platforms.26 Ghosn’s next plan -. Although Ghosn was criticized heavily at first by the Japanese media and government officials. an 8% operating margin. rather than external consultants. This freed up billions in capital for investments in new products and debt servicing. while after NRP. one year ahead of the NRP schedule.

from 27. (Bremner et al. It simply did not function properly at Nissan in the past. raised its stake in Calsonic Kansa Corp.greatly exceeding the industry average of 119. quality problems from this plant (as revealed in J.defying the conventional wisdom that new plants should build existing products to reduce start-up risks. saying “Not everything about the keiretsu was wrong. immediately thereafter. Ghosn held a meeting with Nissan suppliers at which he reportedly changed his position on the value of keiretsu relationships. Ghosn announced a new “project partnership system” in 30 . but at Toyota. In May 2003. from 6th place the previous year. By April 2004. In order to support the high number of new product launches.27 Then in November 2004. we need stronger ties with our suppliers. changes in product and process design followed. and new suppliers -. Power’s initial quality survey of consumers) dragged Nissan down to 11th place. as well as operational problem-solving and extensive worker training.developing 28 new models for release by 2007 -. 2004) Is this quality decline an inevitable consequence of Nissan’s switch to exit mechanisms? Ghosn’s actions in response to this crisis suggest that he is returning to a more collaborative approach. a new workforce. the system seems to be functioning very well.D.including seven completely new products. While the plant itself had start-up problems.6% to 41% (more than a controlling share). Dramatically. Nissan opened a new plant in Canton. With Nissan’s subsidiaries.. a maker of dashboard modules.” Nissan. From now on.. many located the problems in “cheap parts” from suppliers that couldn’t be installed precisely. Soon signs of strain began to emerge. with 147 defects per 100 vehicles-. Mississippi (completed in a speedy 2 ½ years) with three brand-new vehicles. the keiretsu system was too cozy. he flew to Mississippi in May 2004 with over 200 Nissan engineers to undertake extensive examination of quality problems at the Canton plant and at suppliers.

purchasing vice-president Katsuaki Watanabe. How much these collaborations can flourish against the backdrop of Nissan’s extensive shift to exit mechanisms in recent years remains to be seen.which collaborative teams of Nissan and supplier engineers would “review parts from scratch and aim to achieve higher quality at low cost. Often a supplier is asked not only to implement improvements in their own operations but also to help identify where Toyota’s design and manufacturing process for a part increases costs. Many of the savings achieved since the program’s launch in 2000 have come from teams of Toyota engineers working with suppliers on design issues. Nissan may be tempted to resort to exit threats to gain continued price reductions. close collaboration on a project basis. built on the grounds of a former university. Toyota once had 35 different versions of the interior assist grip 31 .” A new plant. will become the next CEO of the company. following its restructuring-driven move away from the voice-without-exit world of the traditional Japanese keiretsu. In one much-publicized example. will take the “supplier park” concept of co-location one step further by putting supplier operations under the same roof with Nissan assembly lines. Nissan recognizes that its suppliers have not been keeping up with the R&D investments made by Toyota’s main suppliers. Nissan’s increased financial stake recognized that “we needed financial assistance to oversee the process from R&D to extending our worldwide production and supply chain. in particular the reduction of part counts and product variants.not least because its architect. According to Calsonic Kansei’s CEO.”28 Thus Nissan appears to be moving back towards “collaboration with trust” as rapidly as possible. as of July 1. Furthermore. and investments in supplier capabilities. Nissan’s new approach emphasizes extensive supplier involvement in design. Toyota’s CCC21 program may also contain indications of where “collaboration with trust” is headed -. 2005.

a small American second-tier supplier.involving careful study and mutual adjustment also how Toyota hopes to cope with the challenge of meeting the “China price” for many components. This type of collaborative problem-solving -.change on Toyota’s side as it did from the supplier. those gains may be less substantial if these new suppliers do not have the strong problem-solving capabilities upon which Toyota relies. sensors. though we predict that they will not switch towards exit despite the potential for shortterm cost gains. and European competitors with extensive manufacturing facilities in China like Robert Bosch and Delphi.providing competitive pressure while also giving its own group of suppliers more time and opportunity to make further improvements.S. now only 3 grip styles cover all of Toyota’s 90 models (Dawson. Stoneridge is a supplier of highly engineered. including control devices. application-specific. The goal will be coming as close as possible to the benchmark price without any sacrifice in quality. Where competing suppliers still have the best price. power distribution. auto companies in recent years. electrical and electronic products to automotive OEMs. After a joint investigation by a CCC21 team and suppliers. 2005). This mix of continuity of relationship with competitive pressure amid growth in demand is characteristic of Toyota’s approach with suppliers.S. This required as much -.if not more -.installed above each door. has coped with the “exit”-oriented supply chain restructuring by U. Toyota may award them a portion of its growing business for that component -. 32 . The prospect that the China price may be so low that it takes considerable time for suppliers to match it will test Toyota’s commitment to this strategy.29 Survival Strategies at Stoneridge: Scrambling to Keep Ahead of Commoditization This case considers how Stoneridge. In any case. CCC21 is identifying global benchmarks for 180 key components on price and quality. including those established by major U.

Brazil. this is identified as necessary but not sufficient because of convergence in performance on this dimension across the industry. Working with partner firms (particularly for joint venture overseas investments) reduces financial 33 . and Asia. It operates 24 manufacturing plants in the U. Stoneridge is pursuing multiple survival strategies. Deep knowledge of the customer is required to support this survival strategy.000 employees and approximately $682 million in annual sales. as is the relentless pursuit of cost reduction described above. Both product differentiation (through features or associated services) and planned product obsolescence are needed to keep a few steps ahead of the threat of commoditization. Pisani argued. Europe. advanced IT. 2) Investments to make (supplier-owned tools. longer payment terms). advanced R&D and prototypes). The product life cycle must be carefully managed. By seizing on technical inflection points as a time to introduce new advanced features. charge-back for quality defects by customer plants). provided these reflections on the demands placed on his firm by the Big Three OEMs during an International Motor Vehicle Program (IMVP) conference session on “Building Tomorrow’s Supplier Capabilities Amid Today’s Pressures. which we have grouped into four categories: 1) Costs to absorb (program management. and 4) Rights to waive (intellectual property rights. 2) 3) Even a small supplier must be mindful of operating in a global industry. Sales to multiple countries and automakers help build the volume that is necessary to amortize product development investments. Mexico.” He enumerated a long list of additional expectations placed on suppliers by their customers.S. 1) The management culture must promote lean production/six sigma thinking in order to achieve operational excellence. Stoneridge’s President and Chief Executive Officer. and has 6. Jerry Pisani.. both in manufacturing but also through involvement in iterated co-design. margins can be boosted on newer products even as prices are being reduced on older products. 3) Liabilities to assume (warranty indemnification. inventory management.and system management components. established contractual provisions). In response. A diversified customer base provides some protection against firm-specific volatility in demand and currency fluctuations.

centralized corporate-level training (more focused and cost-effective than outside training). with joint venture partners. tied to product lines through activity-based costing. 5) Stoneridge’s own suppliers must be carefully managed through disciplines of cost management.risk. and attentive migration of best practices across facilities and divisions. in contrast with only 17% who said they received such assistance from their customers. It is striking that even for a supplier operating in a situation where their customers focus myopically only on price and trust is minimal. These suppliers should be global in their manufacturing footprint but also in their strategic orientation. many of the survival strategies involve collaboration -. not doers). Stoneridge should work to educate them and directly develop their capabilities. 37% responded affirmatively to a statement that “my firm receives technical assistance from competitors”. with an effort to focus on a small group of “best in class” suppliers which whom relationships are long-term. Profitability should be tracked at a micro-level.30 At one level.externally with OEM engineers. 34 . 4) Investments must be carefully monitored to insure adequate returns and to avoid the trap of maximizing revenues at the expense of profits. with their own suppliers. the Stoneridge “survival strategy” seems to point towards developing the most collaborative relationship with customers that can be achieved.and third tier suppliers in the northern Midwest states suggests that even competitors are exchanging information and technical assistance. 6) Learning should be maximized through selective use of consultants (as trainers. Participation in emerging markets provide option value with respect for future demand growth. Inventory turns should be benchmarked against competitors to insure that industry standards are met or exceeded. even as it raises coordination costs/risk. The balance sheet should be kept relatively unleveraged to retain maximum flexibility with regard to future investments. and internally across divisions and within plants. Evidence from Helper and Stanley’s recent survey of small and medium-sized second. Understanding customer needs and differentiating products to offer unique sources of value are two approaches consistent with the “collaboration with trust” mode.

-based “mega-supplier” -. will be to shift effort and contribution (in design insight.of acquiring the knowledge and innovation that they would be losing. Although Pisani did not say so. and that he would be more likely to offer the fruits of Stoneridge’s investments in advanced R&D and product innovation to those customers. The question for the “collaboration without trust” group of customers is whether they will perceive this risk time to remedy it.S. and willingness to forward-invest) to the former group. context where the “exit” mode still prevails. since reverticalization seems highly unlikely -. it is easy to imagine that he would prefer working with customers who are not entirely focused on price.S. these survival strategies are self-protective vis-à-vis customers.Cost Management in the Land of Exit This case study examines the efforts of one U. because it is here that the firm’s intention is to move towards a more collaborative approach is most evident. or whether they will find new sources of suppliers and new means – presumably marketbased. This suggests that one dynamic over time for suppliers who have some customer relationships characterized by trust and some that are not. 35 . The focus of the case study is Delphi’s efforts to shift how it manages its own suppliers. technological innovation.Yet for the most introduce “collaboration with trust” into the U. The only remedy for operating in the face of a low-trust relationship with key customers is to manage product life cycles strictly to keep ahead of commodity pricing pressures. Toyota Comes to Delphi -.Delphi Corporation31 -. Suppliers must be prepared to make quick moves to phase out product lines where pressures are strong to cut prices below actual costs. Financial survival requires avoidance of commitments that build volume and revenue but have low return on investment (ROI).of suppliers withholding contribution and shifting commitments -.

One of Delphi’s primary methods for developing DMS is to hire senior people -.” 36 . a member of the famed OMCD (operational methods and cost management division) -the internal “consulting” group created by Taiichi Ohno and the guardians/stewards of TPS is now a lean sensei at Delphi. and diversify its customer base. Kaz Nakada. We expect them to be committed to being the best supplier in the industry. two of the world’s best lean producers. Delphi also hired Dave Nelson. suppliers. or teacher in Japanese -. I explain that Delphi wants them to be a strategic supplier and is willing to work closely with them. With his help. the margin of both companies will improve. Delphi has hired other Toyota people who are skilled in core TPS disciplines such as cost management and cost profit planning.32 Delphi’s approach to changing supplier relations draws heavily from Toyota and Honda’s approach with their U.Delphi has undertaken a serious effort to implement the Delphi Manufacturing System (DMS) that is closely modeled on the Toyota Production System in order to overcome its problematic cost structure. “We do the meetings one by one. it’s a time to set expectations. where he has a similar role.from Toyota and Honda. I tell them it won’t be easy but they’ll be glad they made the effort. change mindsets.S. I explain how cost management works and that if the process works right.” Nelson starts Delphi’s lean supplier development engineering process with a face-to-face meeting that includes the top executives at the supplier’s location. We ask them to work with us on achieving the best possible understanding of cost and explain that developing this cost standards will take the place of the multiple rounds of bidding they are accustomed to. For the small group of suppliers identified as “core.sensei. who was VP of Purchasing at Honda of America for 10 years. Then we help them become more competitive through lean implementation.

not a tool to use against suppliers by squeezing their profits or sharing their proprietary data with others. developed through exhaustive study of all currently available versions of a component along with a similarly detailed assessment of current and future raw material and commodity prices.” Beginning with “supplier cost” (from the original supplier quote or industry benchmarks).33 Suppliers are told at the start that the remedy to dealing with relentless pressure on margins is achieving ongoing cost reductions. defined as the “lowest total unit cost in the world that is achievable with effort.” Suppliers are told what cost management is not: not a one-time cost-down method. logic. That is set by the “ideal cost” -. Cost management is presented as a “new paradigm” -. The cost standard provides a starting point for reaching the “agreed-upon cost”. Materials from a Delphi lean supplier development engineering session provide insights into how these TPS disciplines foster a “collaboration with trust” approach.” This is achieved through discussion after the customer compares the supplier cost with its book of cost standards. But the “agreed-upon cost” does not define the parameters of the improvement efforts. only very occasionally are multiple bids taken as a check on the cost standards. the former identifies the opportunities for cost reduction while the latter develops supplier capabilities to take advantage of those opportunities. this is also where a short-term cost reduction target is applied. the first goal is “agreed-upon cost. This cost standard takes the place of a multiple bid system.the “modern agriculture” approach that relies on science. This is achieved through a strategic sourcing policy at Delphi that relies on both a cost management methodology and on lean supplier development activities. and facts vs.the ultimate 37 .The next step is to educate the supplier in how the cost management process works. Rather it is an explicitly collaborative pursuit of the “ideal” cost. the old paradigm (“hunters and gatherers”) that relies on “playing poker” through “multiple bidding rounds.

“Reality-based” price reductions closely track cost reductions so that margins are preserved rather than eroded. these cost reduction processes carried out by Delphi together with its suppliers yield savings.” When completely filled out.goal -. This problem-solving approach of working collaboratively towards cost reduction is an archetypal example of interdependent process management. thus reducing direct labor costs. it is crammed with information and symbolic markers of a collaborative effort -. Arbitrary price-downs erode profits because they are mandated independent from progress on cost reduction. Delphi has a process for charting “creative improvement plans. an explicit statement of the Delphi’s “savings sharing percentage” that is applied to each cost savings achieved (making it clear that the supplier retains a percentage as well).names of champions on the customer and supplier sides. Customers help suppliers achieve these cost reductions first by identifying the gap between “supplier cost” and “ideal cost” and then drawing up “creative improvement plans” to apply countermeasures to each identified gap. reinforcing norms of information sharing and transparency. In many cases. The means for achieving the savings varies -. improved 38 .changes in production layout that improve flow and eliminate wasted motion. and very detailed breakdowns of cost savings for every factor of production and for every part number. we have closely analyzed two examples that produced 25% and 32% reductions in supplier unit cost. “Reality-based” price reductions are contrasted with the “arbitrary price-downs” often inflicted upon suppliers by “exit”-oriented OEMs.which takes the minimal cost for each factor achieved anywhere in the world to construct a hypothetical target that is kept as a visible goal to avoid complacency as “agreed-upon” cost targets are reached. improvements in the accuracy of material and cycle time specifications.

design accounts for 5% of costs. following the same logic. If cost management is the short-term. and what remains is the target cost for the vehicle.34 How can design affect so much of cost? Decisions made during design lock in certain requirements for labor and materials and logistics. Hence the primary way to tackle reducing those costs on a long-term basis is to go back to the point in the design process where alternative choices could be considered. 39 . The target cost process developed by Toyota proceeds in the following way. to come up with a target cost for each component. labor for 15%. real-time discipline for achieving cost reduction. vs. Value engineering activities during product development then attempt to achieve that cost target. only 5% for labor. Where does this cost target come from? Unlike cost management. with 30% allocated to indirect and overhead costs. and a lower cost target is assigned. The desired market price for the entire vehicle is set by the OEM based on analyses of consumer perceptions of value. this cost target is calculated following the opposite of a cost-plus logic. 5% for overhead. and then for each design element of the component.and some of the corrections actually increase costs from the original projection by eliminating faulty specifications and erroneous assumptions. In the traditional view. Each aspect of cost is tracked back to that point in the design process when it is locked in. and 20% for materials. materials for 50%. the desired level of profit is subtracted.packaging -. This is then decomposed. In contrast. Cost profit planning views the sources of manufacturing cost differently. cost profit planning is the long-term approach to eliminating cost at the source. which focuses on what is hypothetically attainable from best possible manufacturing practice. in comparison with traditional manufacturing cost accounting. cost profit planning views design decisions as affecting 70% of cost.

but at the same time. recommended a switch from a two-step to a one-step manufacturing process. After starting with a negotiated already-low cost. for its suppliers. consistent with norms of sharing cost savings. they are learning far more about the design aspects than they would under the traditional “black box” system. including the quote from the recognized cost leader. Delphi summarizes.” Further improvements upon the “agreed-upon” cost are achieved through the “creative improvement plan” in the short-term and by cost profit planning – to eliminate costs during the design stage -in the long-term. We saw a summary of one such example that first identified the expected cost from the cost standards set by external research. This dramatic cost reduction was achieved without any adverse impact on the supplier’s profit margins. next showed the range of supplier quotes. suppliers are primarily expected to contribute their input on how to make the manufacturing process easier (what is sometimes called “design-formanufacturing”). This change reduced the part cost 45% below the cost leader’s cost with the old design. This supplier involvement represents a blend of design ideas and operational ideas.The implications for collaborative and iterative co-design are enormous. Prices are set not by multiple rounds of bidding but by comparing the “supplier cost” with the cost standard and having discussions to get to an “agreed-upon cost. costs can be taken out before the manufacturing process even begins. Supplier impact on costs during design can be substantial. and then indicated that the cost leader. 40 . what is different about this approach in the following way. Delphi works with the supplier on cost reductions and price reductions follow as a matter of course. Through supplier involvement with value engineering activities. during the design process. or at least closing the gap substantially. making it much easier to envision reaching the hypothetical ideal cost.

Delphi then tells suppliers what changes in customer and supplier roles are required to make this collaborative mode a reality. 4) Establish “creative improvement plans” with suppliers and document their “reality” -.their actual progress -. Disciplined processes and a focus on facts (“realities”) provide reliable and valid performance outcomes marked by ongoing improvements that build common purpose between the customer and supplier and. from the joint development of creative improvement plans to joint involvement in design to achieve target cost goals: the core value or standard for trust is not performance to preset targets. 3) Identify where the customer’s decisions (design. must: 1) Provide cost breakdown/cost standard information. 2) Discuss operational issues with the customer. 5) Work to help suppliers develop capabilities and pursue ideal cost. trust. An ethic of contribution permeates the customersupplier relationship. It is worth noting the extent to which these Toyota-derived processes of cost management and cost profit planning capture the essential elements of collaborative community as defined in this volume. scheduling. focusing on their manufacturing realities rather than price. 5) Assign target costs to specific design and manufacturing steps and pursue those cost savings aggressively. commodity purchasing) are adding cost and tell the customer. but the willingness and capability to engage in discussion about 41 . and 6) Share cost improvement savings with the supplier. 2) Focus on cost discussions vs. and 6) Share cost improvement savings with the customer. in turn. 4) Take ownership of the creative improvement plan and identify opportunities for cost reduction. over time. The supplier. multiple rounds of bidding and negotiating. 3) Get the lowest upfront (“agreed-upon”) cost through reliance on cost standards and benchmarking. The customer must: 1) Become experts on commodity pricing and manufacturing processes in order to develop effective cost implementing these plans.

We acknowledge to the suppliers that we are in the midst of our own lean transformation. The director of Delphi’s lean production activities told me that in 2003. we are already starting to build the good 42 . There is a strong expectation that all parties will frame problems as opportunities for mutual gain. said “I spend a lot of my time making sure that our purchasing guys act in ways consistent with the relationship we’re building with the supplier through cost management and lean supplier to work towards solutions of problems. As of the end of 2004. This patient one-at-a-time approach seems to be bearing fruit. This number of vendors is greatly reduced from the roughly 9000 suppliers in place when Dave Nelson arrived but is still judged to be too high.000 defective parts per million in 1995 to 28 PPM in 2003. Both sides take responsibility for gathering and sharing relevant information and vow to avoid gaming behavior that exploits information asymmetries.35 This effort to remake Delphi’s relationship with its own suppliers has certainly run into obstacles. manufacturing and purchasing. Nelson has completed 92 expectation-setting meetings with senior executives at this core group. given the intention to have more long-term collaborative relationships. Under the direction of a former Toyota cost management sensei. Delphi saved $500 million from lean-derived improvements in design. Delphi has 4000 suppliers and purchases $14 billion in parts annually. cost standards are steadily being developed for all the Delphi commodities and manufacturing processes. Three years after his arrival at Delphi. The intention is to focus purchases on 1000-1500 suppliers. director of Delphi’s supplier development program. Quality levels have improved from 10. It helps that with the lean supplier development process. Paul Brent. What success has Delphi had implementing this system? The potential impact is certainly huge.

We assess the state of their operations at the beginning and get the benchmarking in place to capture improvements. the difference is roughly 2 to 1. 43 . they can refuse to provide their best knowledge and effort to customers who relentlessly squeeze their margins and violate trust by taking advantage of having access to proprietary supplier information.relationships and trust. privately. 8% lower than those working with other North American OEMs. Delphi provides its suppliers with data on the profits achieved by companies that follow cost profit planning vs. “We’re moving rapidly in this new direction. mostly attributable to the avoidance of time and cost during multiple bidding rounds and to the elimination of design-influenced cost based on supplier input during the product development process.” Evidence of various kinds is emerging to suggest that those following the “collaboration with trust” approach are reaping real economic gains. and then we begin working on creative improvement plans. he said. A customer that insists on bids will end up paying significantly more. those that do not. Suppliers won’t want to do conventional bidding on parts. A 2004 consulting study36 finds evidence from surveys that suppliers working with Toyota and Honda achieve costs that are. We’re working more and more closely with Toyota and getting more business from them. Finally. on average. Once we get past a certain point. in recent interviews. that they are increasingly reserving their most advanced technological innovations for their Japanese customers. We don’t even deal with price. While they can’t afford to refuse business from the Big Three. they are lending us some of their best lean supplier development engineers. it will be impossible to unravel. I see a one-way evolution in the move towards cost management.” When we asked Dave Nelson what he sees in the future. the authors are hearing some suppliers saying.

on the presence of knowledge resources (sensei) and on the growing willingness of suppliers to differentiate among their customers in their level of contribution to design innovation and cost improvement.The Delphi case shows that it is possible for a first-tier supplier with strong historical roots in the traditional U. Delphi’s progress has depended on senior management commitment to putting “lean processes” in place throughout the company. we speculate that the level of discretionary effort applied to making the economic 44 . At the task level.S. (However. we summarize similarities and differences between collaboration with and without trust in Table 7. Implications of Collaboration With and Without Trust This section discusses implications of the above analysis for our understanding of collaborative community.both senior management and the middle ranks of design and manufacturing engineers and purchasing staff . approach of “exit” to move towards “collaboration with trust” in its relationships with both customers and second. Enough people at Delphi have seen the benefits of collaboration with trust -. we see similar reliance on iterative co-design and interdependent process management regardless of the level of trust at the governance sustain the commitment to this approach. We address the dynamics of collaboration in this context by addressing four questions: 1) Is trust needed for collaboration? 2) Can the gains from collaboration with trust be appropriated by other firms that are practicing collaboration without trust? 3) What are the micro-level social and economic dynamics of collaboration without trust? 4) What constraints are faced over time by collaboration with trust? Before doing so. The difference in legacy mode primarily affects the governance level or purchasing regime.and third-tier suppliers.

relationship work effectively (including solving emergent problems) may differ even at a task level. Collaboration without trust can emerge where “exit” approaches to purchasing routines and governance mechanisms persist. we believe that empirical investigation will be needed to determine the conditions under which this is possible. However. This work requires collaboration by necessity. We now address a series of questions on the dynamics of collaboration with and without trust. with respect to trust at the task level.) <<Table 7 about here>> With respect to information exchange. collaborative co-design – when wellexecuted – will lead to superior outcomes and provide experiences of success that will help build 45 . particularly given a (still) largely integral product architecture and the diffusion of innovations of iterated co-design and shop-floor problem-solving associated with lean production (Toyota Production System). we predict that collaboration will prove to be more successful if it takes forms that allow for the emergence of trust over time. Finally. as a consequence of a “trickle-down” effect of the lack of trust at the governance level. We argue that a shift of knowledge-intensive design work from OEMs to suppliers is underway.something that should have long-term consequences for quality. Given a volatile set of external circumstances and ongoing advances in technology that must be absorbed. as part of a move to deverticalize the automotive industry. we see similarly high amounts at the task level but a striking feature of low trust at the governance level is a constricted and formalized flow of information -. Is Trust Needed for Collaboration? The evidence of this chapter suggests that trust is not a precondition for collaboration. The evidence of this chapter is that firms are being pushed in the direction of collaboration whether or not they have prior traditions of trust-based supplier relations.

Customer B could demand to receive the same annual cost reduction offered to Customer A. there is a risk that other customers might “free ride. Now suppose that Customer B..S. Can the gains from collaboration with trust be appropriated? Since.a “rising tide lifts all boats” hypothesis. prompted by the Japanese transplants. under the assumption that the supplier capabilities will diffuse naturally within the firm. Or.” Imagine that a supplier has improved its design and manufacturing capabilities through collaborative interaction with Customer A. increasingly. some of which help the supplier improve through supplier development activities. suppliers work with multiple customers. one dedicated to a Big Three customer and the other to 46 . easy to destroy with a single exploitative act during the purchasing process. It may also exist primarily at a microlevel of personal relationships among engineers and managers who have worked together during co-design and may not diffuse readily to individuals who are not part of the co-design experience. Two production lines side by side within the same supplier plant. have given their Big Three competitors a considerable upgrading in cost and quality -. not willing to make such investments in supplier capabilities and operating under a low-trust purchasing regime. Many observers do make an argument of this kind. This trust is likely to be fragile at first. claiming that the improvement of supplier capabilities in the can gain the benefits of this supplier’s superior capabilities simply by awarding it some business. figuring that the only way the supplier could attain these reductions would be through the same process improvements. Yet Dyer’s (1997) evidence suggests that this is wrong: he found that the diffusion of best practices within the same supplier can be remarkably low. such as the customer sending the supplier’s proprietary design information to competitors seeking a competitive bid.

but the automaker purchasing agents are so intent on squeezing suppliers for every last nickel that they are eliminating the prospect of effective collaboration. For example. because GM mandated the use of standard containers. a supplier drawing on deep knowledge of one customer achieved through a long-term relationship governed by trust and frequent interactions will be able to provide more innovative designs and insightful process improvements than a supplier that is dutifully applying generic process management routines. For example. consider a general hypothesis that the engineers want to collaborate. Toyota worked with a supplier plant to implement one-piece flow of parts through the production process. Unlike quality control/improvement. we would speculate that value and personality differences are important: that an “ethic of contribution” among engineers inclines them more positively towards collaborative activity than purchasing agents. The consequences over time will be a growing gap between what Customer A and Customer B are receiving. These techniques were not useful to the GM line in the same plant. which is often a win-win situation between workers 47 . In the absence of direct evidence. others are relationship-and context-specific. containers that were so large that forklifts were required to move them. Our view is that while some capabilities are generic and all customers will benefit from a supplier’s improvement. had widely different performance. What are the micro-dynamics of collaboration without trust? One way to describe the micro-level motivations and dynamics of collaboration without trust is by focusing on the typical culture and experiences of different professions. Even without deliberate intent.Japanese transplant. emphasizing the use of small containers and short lead-times. His explanation was partly based on organizational barriers to diffusion of knowledge and partly on differences in production philosophy that affected manufacturing outcomes. even from the very same supplier.

causing them to evaluate the benefits of collaboration (vs. trust-undermining price-reducing actions) in each period for which they are rewarded. purchasing agents are judged based on their ability to contribute to these tasks. Second. giving them an incentive to think about the value of collaboration over a longer period. This decision to focus on the “reality” faced by the supplier’s manufacturing plant.” Engineers are evaluated on completion of interdependent tasks (such as whether designs are finished on time and pass crash tests) that require collaboration between counterpart engineers at automaker and supplier. it provides a different set of goals and experiences for purchasing agents that over time changes their culture. and hence to reduce prices only after reducing costs. parts sourcing may lack a task that would pull automaker purchasing agents and supplier salespeople into greater interdependence and help develop norms of reciprocity. The Toyota approach involves coming to an agreement upon an initial cost based on careful preparation of cost standards and then pursuing cost reduction to reach the target cost. First.and managers in a manufacturing setting.37 What constraints are faced over time by collaboration with trust? 48 . Another framing would use the language of incentives: “Engineers and automaker purchasing agents are each responding rationally to the incentives they face. provides the basis for norms of mutual gain and reciprocity and builds trust. Both framings contribute to an explanation of why the cost management/ cost profit planning approach used by Toyota (and being adopted by Delphi) can be successful at the governance level. Automaker purchasing agents are rewarded for reducing piece prices. facilitating collaboration with trust. They thus ignore a potential lost stream of future benefits from collaboration.

are approached according to collaborative norms of information-sharing. ‘free’ design work absent any contract) with so few safeguards. firms following these norms should possess a greater adaptability than firms who are able to make very quick short-term decisions. Meta-routines for examining skeptically all current practices may be important to avoid stagnation. However. This would suggest a steady evolution towards collaboration with trust as the dominant mode. and also predict that suppliers will prefer to deal with customers operating collaboratively with trust.Long-term collaborative relationships with little or no threat of exit can certainly face constraints in the form of complacency. suppliers have assets they 49 . reciprocity of effort. Yet we do not expect that collaboration without trust will disappear any time soon. Furthermore. However. “groupthink” tendencies. we predict that collaboration without trust will not perform as well as collaboration with trust. shared distribution of gains and losses.whether from external or internal sources -. etc. It is for this reason that MacDuffie (1995) describes the Toyota Production System in terms of “flexible production” rather than “lean production. as long as all unexpected shocks -. Predictions about the future of “collaboration without trust” Above. From the perspective of transaction cost analysis with which we began this chapter.g. sudden environmental changes or financial pressures may require rapid response that might not allow fulfillment of the commitments of collaboration –a tension between flexibility and the market as mentioned early in the volume. rigidity. it may seem surprising that suppliers would make investments in assets tied to one customer (e.” This perspective is supported by empirical evidence that flexible production plants can achieve high levels of product variety without suffering a decrement in cost or quality performance.

need to pay off. and can bid commodity producers down to their marginal costs. To the extent that tasks relatively predictable in their frequency. It is 50 . in plant and equipment.towards alternate fuels. new drive train designs such as fuel cells. new uses of information technology within the vehicle -. with more market-based modes of module procurement. we predict that collaborative activity between automakers and suppliers will remain high and might even increase.break the dominant design sufficiently for a more modular product architecture to become possible. the total requirement for collaboration may decrease. Customers know this. collaborative activity could shift towards module definition and interface standardization and away from high interdependence during design.. they have bargaining power over noncommodity producers as well. The overcapacity situation affecting the parts sector is not likely to change quickly. indeed. the forces promoting the addition of capacity in low-wage countries (their large direct-labor cost advantage) as well as those slowing the removal of existing capacity in high-wage countries (the long-lived nature of capital equipment) remain strong.would increase. and if they can cover part of these costs. timing. Even in the presence of dramatic technological change. and informational requirements can be automated (relying on sophisticated software that can be contextually responsive). technology etc. new materials. they are still better off than they would be without any sales. To the extent they are willing to accept less than perfect substitutes for their first-choice product. even as the proportion of non-automated tasks that are collaborative – and that would benefit from trust -. This transition could last for a long time. if fundamental technological shifts -. It is also possible that some future trends affecting supply chains in the automotive industry could decrease the necessity for collaboration. Furthermore.

tempting to speculate that customers would benefit from relationships with less commitment when “radical” or “disruptive” innovations (Abernathy. but does not in practice seem to have been the case in the auto industry. 51 . etc. and overwhelming product and process complexity generate such a challenging and diverse set of decisions about purchasing strategy. But it is not clear that automakers using more collaboration were slower than others. What we can assert with confidence is that the automotive industry will continue to be a fruitful context for examining modes of inter-firm economic exchange. The important intangibles of “look and feel” associated with automotive products and brands will resist any sweeping move towards modularization and preserve the knowledge-intensity and idiosyncratic requirements for systemic integrity that place collaboration at the heart of this industry. since its global scale. giving existing suppliers a chance to learn about (or buy) the new technology. suspension. 1978. technological scope. Christensen. high volumes and high demand for reliability) that mean even a new supplier of a radically new product must take a lot of time to learn about unchanging features of the industry. Perhaps this continuity is due to features of the auto industry (such as the integrality of its products. This seems plausible. and experimentation with alternative power trains). and in fact most of the new technologies are supplied in partnership with existing suppliers. The industry has over the last twenty years adopted a number of such innovations (such as the change from mechanical to electrical control of carburetion. 1997) make their current suppliers’ capabilities obsolete.

and Management in the World Auto Industry. Brace & World. Strategy and Structure: Chapters in the History of the American Industrial Enterprise. 4: p. Organizational Coupling. 7(6): 649-666. William J. 50. Chandler. “Knowledge Specialization. and Chester Dawson (2004). Jeffrey H.(2000). Design Rules: Volume 1. Baldwin. Dyer. Fixson. Dyer. Boston. Harvard Business School. Dominance. vol. and Clark. The Innovator's Dilemma. Brian. MA: Harvard Business School Press. Jeffrey (2000).” Organization Science. Dyer.References Abernathy. Carliss Y. MA: MIT Press. Liker (forthcoming). Stefano. Chesborough. Organization.S.. K.. T. Y. Ro. Automotive Cockpit Industry. Jeffrey H. Charles (1999). February 21: pp. Robert (1984). “A ‘China’ Price for Toyota. (1978). 4(16): 386-405. Chester (2005). 5051. Fine. A. Christensen. Ronald (1937). Boston: Harvard Business School Press. vol.” Business Week. “The Relational View: Cooperative Strategy and Sources of Interorganizational Competitive Advantage. "Modularization and Outsourcing: A Study of Generational Sequences in the U. Alfred D. Clark. (1997). Oct.” Academy of Management Review. Baltimore: John Hopkins University Press. and Harbir Singh (1998). and the Automobile Industry.” Administrative Sciences Quarterly. New York. ed. Brusoni. The Productivity Dilemma: Roadblock to Innovation in the Automobile Industry. Cambridge. Kim B. The Evolution of Cooperation. MA: Harvard Business School Press. S. Axelrod. The Power of Modularity. and Death. and J. (1997). Product Development Performance: Strategy. Collaborative Advantage: Winning Through Extended Enterprise Supply Networks. Reading. New York: Harcourt. Coase.” Business Week. General Motors. and Fujimoto. "The Nature of the Firm. Clockspeed. “Does Governance Matter? Keiretsu Alliances and Asset Specificity as Sources of Japanese Competitive Advantage. Dawson. “Towards a Dynamics of Standards: Their Emergence. 46(4): 597-621. MA: Perseus Books. Clayton M. Henry (2002).." International Journal of Automotive Technology and Management. vol. and Keith Pavitt (2001). “Nissan’s Boss. Boston. 52 . B. (1991). Andrea Principe.” working paper. Bremner. 23( 4): 660-679.(1964). and the Boundaries of the Firm. Chandler.D. MA: MIT Press." Economica. vol. Giant Enterprise: Ford. (1962). Cambridge. Gail Edmondson. Basic Books.

Gary (2004). Susan. Sebastian (2005). Albert (1970).S. The Evolution of the Toyota Manufacturing System. “The Shifting Boundaries of the Firm: A Case for Information. and Modularity Conditions. Helper. MA.” Business History Review. Carlos (2002). 53 . vol. MIT.S. Helper. Susan and Marcus Stanley (2004).” in Bruce Kogut. 80(1): 37-45. DC: Brookings Institution Press. Susan and John Paul MacDuffie (2003). 25(2): 39-51. Italy.. vol. and Supply Chain Design Decisions.” PhD Dissertation. London Business School.). vol. “Pragmatic Collaborations: Advancing Knowledge While Controlling Opportunism. 10(3): 443-83. "Industrial Clusters. New York: Currency Publishing. 11(1&2): 89-107. Hirschman. R. Susan and Janet Kiehl (2004). “Long-Term Supplier Relations and Product-Market Structure. Jacobides. 331-380. pp. 36(3): 77-84. “Supplier Relations in Japan and the United States: Are They Converging?” Sloan Management Review. Cambridge: Harvard University Press. vol. Process. Litan and A." conference paper for Clusters. Ghosn. Cambridge.” Industrial and Corporate Change. Helper. Social Capital. vol. Fujimoto. 23(3-4): 345-369. Helper. vol. Susan (1991). vol. Stephen (1991).” Business and Economic History (best papers volume). vol. Helper. MIT Press.” Industry and Innovation. Shift: Inside Nissan’s Historic Revival.Fixson. Hochfelder. and Charles Sabel (2000). Ghosn. UK: Oxford University Press. Susan.” Journal of Operations Management. “Suppliers and Intermediaries. “Product Architecture Assessment: A Tool to Link Product. Component Manufacturing Industry.” working paper. pp. ed.” Harvard Business Review. and International Competition in the U.” Industry and Innovation. "Strategy and Irreversibility in Supplier Relations: The Case of the US Automobile Industry. 11(1&2): 45-79. “Emerging Strategies and Forms of Governance in High-Wage Component Manufacturing Regions. “Joint Product Development in the Early American Auto Industry. Susan and John Paul MacDuffie (2001). August 4. 178-213. Oxford. Rivlin (eds. “E-volving the Auto Industry: E-Business Effects on Consumer and Supplier Relationships. Carlos (2004). Susan and Mari Sako (1995). Coordination. Michael (2002). David and Susan Helper (1996). Modena. The Global Internet Economy. John Paul MacDuffie. “Towards a Cooperative Commonwealth?: Labor and Restructuring in the U. Helper. Herrigel.” Journal of Law Economics and Organization. Helper. and Canadian Auto Industries.” in Tracking a Transformation: Ebusiness and the Terms of Competition in Industries. Herzenberg. Industrial Districts and Firms: The Challenge of Globalization. Helper. 8(3): 561-581. Exit. and Loyalty. Takahiro (1999). Washington. “Saving the Business Without Losing the Company. “Developing Supplier Capabilities: Market and NonMarket Approaches. and David Levine (1992). vol. Voice. 65(4): 781-824.

” Journal of Economic Behavior and Organization. John Paul and Susan Helper (1998). “Modularity in Technology and Organization. Cambridge. and the Antimonopoly Law in Japan. “Nissan Lifts Oversight in U. B. S. (2001). Fruin. vol. Appropriable Rents.” in J. John Paul (1995). 54 . Jan. “Creating Lean Suppliers: Diffusing Lean Production Through the Supply Chain. Mackintosh. 68(3):79-91. Prahalad.. Sako.” Enterprise and Society: The International Journal of Business History. International Motor Vehicle Program. Oxford. Oxford: Oxford University Press. New York: McGraw-Hill. vol.). Crawford. vol. vol. “Neither Modularity or Relational Contracting: Inter-Firm Collaboration in the New Economy.” Journal of Economic Literature. Richard N. Mari (1992). and R.S. Sept. vol. “The Global Leadership of Carlos Ghosn at Nissan. 40(3): 952-953.” Harvard Business Review. 48 (2): 197-221. pp. Adler (eds. W. “The Core Competence of the Corporation. and P. Oxford. Liker. 49(1): 19-37. Making and Selling Cars: Innovation and Change in the U. 28.” Nikkei Weekly. (2002). 5(3):388-403.” Management Science. Sabel. vol. Strategic Industrial Sourcing: The Japanese Advantage. R.S. James M. Tadahisa and Kazuyuk Okudaira (2005). and Trust: Inter-Firm Relations in Britain and Japan.K..” Journal of Law and Economics. UK: Cambridge University Press. Toshihiro (1994). John P. Nishiguchi. 43(4): 479-502. Liker. UK: Oxford University Press. Jeffrey (2004).S. vol. 47(1): 121-137.M. Alchian (1978).Kenney. “Modularity and the Automobile: What Happens When the Concept Hits the Road. 24. Baltimore: Johns Hopkins University Press. Langlois. MacDuffie.. James (2004).K. “Vertical Integration. 21(2): 297-326.” Financial Times. MacDuffie. Florida (1993). this volume Sabel. Klein. vol. John Paul (1997). Rubenstein.” Industrial and Labor Relations Review. and Jonathan Zeitlin (2004). UK: Oxford University Press. The Toyota Way. Trade Associations. Prices. John Paul (2005). Beyond Mass Production: The Japanese Lean System and Its Transfer to the U. M. “The Road to Root Cause: Shop-Floor Problem-Solving at Three Automotive Assembly Plants.” Thunderbird International Business Review. and Gary Hamel (1990). MacDuffie. Millikin.” working paper. Mari (2002). Remade in America: Transplanting and Transforming Japanese Management Systems. Automotive Industry. Mikawa. MacDuffie. Charles F.. “Human Resource Bundles and Manufacturing Performance: Organizational Logic and Flexible Production Systems. “Cooperative Capitalism: Self-Regulation. Charles F. 30: p. 154-202. and Dean Fu (2005). C. and the Competitive Contracting Process. “Ghosn Recasting Keiretsu Suppliers. A. and A. Quality. Sako.

” presentation at Japan Auto Industry Symposium and Sponsors Meeting (co-hosted by IMVP.html (forthcoming in Industrial and Corporate Change).” Research Policy. 55 .” presentation at IMVP Spring Researchers Meeting. The Automobile Industry Since 1945.. Akira (2003). 116(12): 26. “Structural Changes of Japanese Supplier System. Sept. Andrea Principe. Whitford. “Nissan: The Squeaks Get Louder. International Motor Vehicle Program. Lawrence J. Cambridge. Welch. New York: Free Press. Sturgeon. and Michael Hobday. Markets and Hierarchies. May 17: p. and the Prospects for Inter-firm Collaboration in U. Tokyo. June 2-3. Takeishi. Philadelphia. (2004). James P. “Structural Changes in Japanese Supplier System.” Business Week. The Economic Institutions of Capitalism. pp.. and Toyota.S. “Modular Production Networks: A New American Model of Industrial Organization. Karl (1995). 12. MA: Harvard University Press. vol. Oliver (1985). Christopher A. New York: Free Press. Josh and Jonathan Zeitlin (2004). Takeishi.” Automotive Design and Production. The Machine That Changed the World. Oxford. Womack.Sako. http://imvp. Timothy (2002). “Governing Decentralized Production: Institutions. Akira and Yoshihisa Noro (2005).” Industry and Innovation. (1971). “Modularity and Outsourcing: The Nature of Co-Evolution of Product Architecture and Organizational Architecture in the Global Automotive Industry. Public Policy. UK: Oxford University Press. eds. Hosei Univ. 11(1&2): 11-44. Nissan. and REITI). 11(3): 451-96. New York: Rawson Associates.” in The Business of Systems Integration. vol.” Industrial and Corporate Change. Ulrich. Japan. Sawyer. Williamson.” working paper. Andrew Davies. David (2004). 44. vol. PA. “The Role of Product Architecture in the Manufacturing Firm. 229-53. Oliver (1975). Daniel Jones and Daniel Roos (1990). Manufacturing. White. 24(3): 419-440. Williamson. “Supplier Development at Honda. Sako. Mari (2004). Mari (2003). “Speed and

Estimates. Big 3 Company General Motors Ford Chrysler 1975 75% 66% 50% 1995 66% 50% 33% 2005e b c d a. sales dependency. d. Delphi Corporation. Based on IRC data on component transactions for 200 key components. Figures in red are purchases from in-house (vertically integrated) parts suppliers. Chrysler was purchased by Daimler-Benz in 1999.around 30% -.after Delphi and Visteon spinoffs. Ford. 56 .Table 1 Vertical Integrationa -. as defined by IRC based on financial affiliation. Rubenstein (2001) estimates that GM. includes bought-in materials for in-house produced parts (Nishiguchi. with 88% of its sales to Ford (70% in 2004). supplemented by Nobeoka and Manabe survey on nature of components. b. Visteon Corporation. with worldwide sales of $27 billion ($21 billion in N. when spun-off from Ford in 2000. was immediately the world’s largest supplier. Restructuring of Nissan’s keiretsu began in 2000.A. analyzed by Akira Takeishi and Yoshihisa Noro. when spun-off from GM in 1999. Figures in green are purchases from keiretsu suppliers. These time series data reflect 72 components that are the same for the entire period. e. Nissan figures in 1999 are 7% in-house and 55% keiretsu suppliers. c.) 78% of its sales were to GM (54% in 2004) c. d.S. 1994. Table 2 Vertical Integrationa -. Rubenstein.U.Japan Company 1990 1996 2002 1984 12%/60% 11%/62% 10%/62% 10%/62% Toyota Nissand Honda 10%/56% 5%/39% 10%/53% 5%/46% 8%/56% 5%/48% 4%/36% 4%/41% a. and Chrysler levels of vertical integration were similar -. 2004. for car and light truck manufacturing only. had worldwide sales of $12 billion. reflecting the new strategy after the alliance with Renault. and historical relations. 2001) b. a few parts plants have been sold or closed.

327 billion 9) Faurecia 10) Siemens VDO Auto. $24.508 billion $13.Table 3 The Rapid Increase in Automotive “Mega-Suppliers” # of suppliers with: 1992 >$10bn 3 global sales $5-10bn global sales $2-5bn global sales 2 11 1995 3 11 36 1997 4 10 33 2000 8 10 35 2004 11 12 41 Source: Automotive News.937 billion $19.600 billion 57 . $11. Table 4 TOP TEN GLOBAL OEM PARTS SUPPLIERS . America $19.104 billion 3) Magna International $19.500 billion 5) Johnson Controls 6) Visteon Corp. $17. Corp.927 billion 4) Denso Internat’l.200 billion 2) Delphi Corp. of America $15.000 billion 7) Lear Corp.700 billion $17. 8) Aisin World Corp.2004 1) Robert Bosch $27.

Table 5 From ‘Exit’ and ‘Voice’ to Hybrid Collaborative Mode of Exchange EXIT Arms-length & transactional Open for new suppliers to bid Competitive selection by low bid – frequent and speedy exit Design simplified by customer to enlarge pool of suppliers No equity stake Contracts for governance Codified procedures VOICE Long-term & relational Set of potential suppliers mostly closed Selection based on capabilities -.exit rare and slow Design controlled by customer. supplier involved via resident engineer Often an equity stake Norms /dialogue for governance Tacit procedures HYBRID -. after a vetting period Competitive assessment -intermediate frequency and speed of exit Larger design role for supplier.NEW COLLABORATIVE Long-term & relational Open to new suppliers. attention to supplier design capabilities Equity stake depends on criticality of technology Norms + process management routines for governance Process management routines make procedures explicit Figure 1 Two Patterns Emerging from Hybrid Collaborative Mode EXIT VOICE Hybrid Collaborative Collaboration Without Trust Collaboration With Trust 58 .

Figure 2 Source: Delphi Corporation 59 .

Delphi MUST shift the focus of our efforts. 70% % Influence on Total Cost 20% 5% 5% Courtesy of Munro & Associates % ign 5 Des 0% ial 5 ater M ting c oun t Ac Cos onal bility tura a d i ti Tr ufac M an 30% rden % Bu or 15 Lab 60 .Figure 3 2 Since DESIGN impacts 70% of the Total Cost.

3% – 1 operator running 3 CNC stations 8.1% AFTER COST – No manual stacking required – U-shaped module – Improved communication – 1 piece flow – 6 operators – 126 pieces/hour 8.33 Euros Cost Leader and Suppliers Recommend 1-Piece 1-Piece Design in Process 0.28 – 0.9% – No communication – 5 piece flow – 6.5 operators – 90 pieces/hour 25% reduction in overall cost with improvement to supplier margins Source: Delphi Corporation Table 6b Cost / Profit Planning Example Product: Project: Division: Annual Volume SOP: Component: Design: Costbook: Supplier Quotes: Cost Leader Estimate: VE Activities: Status: New Cost Leader Estimate: Source: Delphi Corporation Oil Control Valve OCV E&C 2 million + November 2005 Frame 2-Piece Stamping with Spot Welding 0.75 Euros 0.Table 6a Cost Management—Metallic Assembly BEFORE Machining – 1 operator running 1 CNC station Part Stacking – 1 operator running 1 stacker – Manual stacking performed as needed Assembly – Straight line layout – 1 operator running 2 stackers 5.18 Euros (45% Avoidance Over 2-Piece) 61 .80 Euros 0.

? task Emergent High 62 .Table 7 Collaboration With and Without Trust Collaboration Without Trust Legacy Mode of Exchange Exit Collaboration With Trust Voice Long-term and relational • Iterative co-design • Interdependent process management • High discretionary effort (?) Governance Level Adversarial/short-term & (Purchasing Regime) arms-length Task Level (Design/ Engineering) • Iterative co-design • Interdependent process management • Low discretionary effort (?) Information Exchange Trust Low at governance level. High at governance high at task level and task level Low governance.

7 8 The Clark and Fujimoto typology isn’t comprehensive from the perspective of this chapter. Gradually. continuous improvement routines that pursue “root cause” solutions to cost and quality problems through rigorous. Ironically. For example. US suppliers played a key role in product design and innovation in the industry. which focused on direct labor costs and did not capture other kinds of costs as effectively. Attention to these differentials were magnified due to automakers’ accounting systems. See Fujimoto (1999). “loyalty”. “atrophy the development of voice”. in the supply chain context. White 1971. is that a supplier will benefit when its customer displays some loyalty and pursues voice mechanisms. MacDuffie. the possibility of exit may. in Hirschman’s words. there is an increased likelihood they will resort to “voice”. vertically integrated divisions came to take on this role. 1996. and suppliers (except for a few large independents) played the role of ‘cheap capacity’. the OEM could essentially say to the supplier “here are the 63 . such as a lower price) seem to outweigh the loss of the on-going relationship (which entails costs that are often hard to measure. 2 1 Hirschman offers a third category. comprehensive experimentation with different manufacturing approaches. a supplier with small stamping presses might want to design small parts and weld them together. such as loss of trust). Helper 1991. since especially at the beginning of relationships. MacDuffie (1997).Endnotes Collaboration between automakers and their suppliers was not invented by the Japanese. larger parts. which affects the relationship between exit and voice by discouraging those who are dissatisfied with an economic relationship from leaving it immediately. Helper. Vertical integration fell in the 1930s. or whether they were rationalizations of decisions taken due to labor cost differentials. as they strived to provide onestop shopping for their de-verticalized customers. the mega-suppliers often became more vertically integrated. choosing to display such loyalty is central to the strategy associated with this mode of exchange. 4 5 6 It is unclear how much of an independent effect the new theories had. See Hochfelder and Helper. Womack et al. and Sabel (2000). 3 We define both ‘collaboration’ and ‘trust’ below. and just-in-time inventory practices that interlink supplier and automaker factories in a minimal-buffer regime that requires precise coordination and well-developed logistics capabilities. The implication. while a supplier with larger presses would prefer to produce fewer. See also Helper and Stanley (2004). Until the 1920s. thus providing valuable information to improve the relationship. These practices include: simultaneous engineering during product development. (1990). With less likelihood of “exit”. there are often bumpy periods in which the benefits of switching (often tangible. Indeed. This addresses the situation in which the presence of both exit and voice options yields a net bias towards exit. Black box subcontracting could be purely market-based. involving iterated communication about a design space that is steadily narrowed to reflect emergent understandings of which choices will optimize the design across multiple dimensions of performance. Susan Helper coined this phrase in connection with a presentation at an International Motor Vehicle Program (IMVP) conference in September 2002. then rose in the 1950s before falling again in the 1980s and 1990s.

we mean undertaking tasks that require on-going discussion. See Langlois. and/or reward structures not changing to reflect the changed strategy. the outcome of these discussions is frequently a contingent action not anticipated or provided for in a contract. 2002. 9 See Ulrich (1995) for the classic definition of modules as involving a one-to-one mapping from component (design element) to function. 2005. in a non-collaborative arm’s-length or ‘exit’ mode.autonews. identifying design problems. with the expectation that one will not be taken advantage of. this approach is what some observers imagine will be commonplace with a modular product architecture. 2001.parameters. see MacDuffie (2005). see Fixson. subjective requisites of voice (nurture suppliers. meet them. Jacobides. but don’t accept substandard performance). we would like to understand better the extent to which this slow reaction is due to some inherent resistance to change. For a useful reconceptualization of product architecture that challenges and extends the Ulrich definition. track the evolution of cockpit designs over three periods while also tracking the extent of outsourcing of design and manufacturing tasks from automakers to suppliers. 2002 issue of Automotive News. 2002. This definition corresponds to Sako’s (1992) definition of ‘goodwill’ trust. For an expanded treatment of this line of argument. 64 . Certainly the issue of how extensively the parties work together on determining parameters. and jointly solving them is central to the issue of collaboration in supply chains. indeed. as discussed below. and Liker. By trust. This transition seems particularly difficult when a firm moves from the relatively clear and measurable incentives of exit (minimize piece prices and the costs of switching suppliers) to the more multi-faceted. Sabel and forthcoming.thefabricator. 14 By ‘collaboration’. They find that outsourcing is much more extensive than any change in product architecture. For the case of purchasing. the category of “black box” in Clark and Fujimoto’s data collection did involve considerable interactive deliberation between OEM and suppliers over the parameters of a component and how they were to be achieved. Ro.” Indeed. Top management frequently expresses frustration that their subordinates do not react quickly to changes in strategy.cms?articleId=40392. the entire text is available at http://www. In future research we will investigate the relationship between changing incentives and changing attitudes. In practice. 10 11 12 13 A similarly incorrect prediction of an inevitable decline in collaboration due to increasing standardization was made by proponents of e-business in the late 1990s. Brusoni et al. 16 17 18 15 a division of labor is agreed on in advance and little discussion occurs across the agreed-upon interface. 2004. we mean taking actions that leave oneself vulnerable to another party. In contrast. 2002. 2004. Chesborough. 2001. a difficulty in changing habits. if you don’t we’ll go somewhere else. This speech was reported in the August 12.cfm?ID=25 Herrigel. Whitford and Zeitlin. Fixson. the changes described as “stage 1 modularity” are primarily to allow one-step installation during assembly and do not reflect any of the more fundamental attributes of modularity. See Helper and MacDuffie. Perhaps a closer concept to our description of “pragmatic collaboration” is relational subcontracting (the term used in transaction cost economics).

2001. Toyota Motor Corporation. 22 23 Ohba seminar. 30 31 Delphi is a world leader in mobile electronics and transportation components and systems technology. Prencipe. See Brusoni.Dynamics.. a reduction of approximately one-third in procurement costs.provide comprehensive product solutions to complex customer needs. Propulsion. 2005). and Pavitt (2001). This was not an acquisition but an alliance. 25 26 Ghosn’s account is captured in various interviews and in a Harvard Business Review article (Ghosn. Toyota has largely succeeded in reaching this goal. 2005. 2003. Mackintosh (2004). 1997. Fortune. 24 This account of the Nissan turnaround is based primarily on Milliken and Fu (2005). See Mikawa and Okudaira (2005). cited in Sako. Delphi has approximately 184. 42 joint ventures. the Nissan board would continue to choose the CEO. See Welch (2004). the exit-based legacies of documentation and codification (plus lower quality standards) may mean that a firm like GM can more quickly incorporate new suppliers with weak management skills.delphi.7 billion in cost savings for the fiscal year ending in March 2005 has not been met. however.000 employees and operates 167 wholly owned manufacturing sites. See Chester Dawson. Bremner et al (2004). In response. USA. and Sabel. Paris. saving nearly $10 billion over its five-year span. By most accounts. Thermal & Interior Sector and Electrical. Business Week. The target of $1. 27 28 29 In contrast. 2004. 2000. Delphi's two business sectors . Delphi can be found on the Internet at www. Nissan retained its name and board of directors. Multi-national Delphi conducts its business operations through various subsidiaries and has headquarters in Troy.19 20 21 See Helper. supplemented by articles from Business Week. 2002). Helper and MacDuffie. with an extended treatment of the Nissan turnaround in a book titled Shift: Inside Nissan’s Historic Revival (Ghosn. Electronics & Safety Sector . February 21. See Sako (2002). MacDuffie. Tokyo and São Paulo. See also Liker. 53 customer centers and sales offices and 33 technical centers in 40 countries. Mich. Sawyer (2004). Institute for International Economic Studies Seminar Series #9706. The supplier development program that Nelson directed at Honda is analyzed in detail in MacDuffie and Helper (1998). Toyota is aiming to reduce the number of steel parts in an average vehicle from 610 to 500. Forbes. falling short by 15% because of high steel prices. 32 65 .com (as of May 13. See Helper and Stanley (2004). Brazil. 2004). and Nissan would take the primary responsibility for implementing a revival plan.

etc. engineers may see this payoff increase due to greater familiarity with each other’s design practices. industry average. followed closely by Honda. which is possible due to increased bargaining power by OEM's due to supplier excess capacity. MacDuffie. Planning Perspectives. 2005. Or. in these relationships suppliers a) pay for things that OEM's used to pay for (such as program management). The gap between Toyota and GM has widened each year from 2002 to 2004. but purchasing agents often have no way of valuing this familiarity.33 This account is drawn from presentation materials from a Delphi meeting with its suppliers in November 2004. they invest sometimes millions of dollars to create (to take a real example from our interviews) a complete design for the interior of a minivan. the purchasing agent might have a longer discount rate than most.. there might be uncertainty about the buyer’s ability to hold up the supplier. This scenario also differs from that of Axelrod (1984) and of Helper. Suppliers might make investments in “non-redeployable assets” (Williamson. a design that was not used because the automaker decided not to make that particular model. without an increase in the piece price they receive and b) make specific investments in design without perfect safeguards (to use Williamsonian language). this isn't always true--the OEM might not have an alternative supplier that can step in immediately. 34 35 36 2004 Global Supply Management Survey. As we have argued. January 10. in transaction cost theory. Suppliers accept these new terms because they have sunk investments. In our case. When suppliers rank OEM-supplier working relationships. These investments (eg in design engineer time) might allow them to earn a return on other assets (a factory). if a firm invests in a specific asset without contractual safeguards. Toyota is first-ranked. 37 66 . 1975) (case b) for a variety of reasons. cited in Delphi presentation to suppliers. the new price is still greater than their marginal cost. Suppliers’ willingess to take on new variable costs (as in (a)). Sabel (2000) in that there is not necessarily a increasing payoff to collaboration as the number of periods of collaboration increases. Why do suppliers continue to participate in collaboration without trust? As we have discussed. or other factors might make the net present value of cooperation greater than that of defection. there is a 100% probability that it will not recoup the fixed cost of this investment. is a disguised price cut. December 2003 and January 2005. That is. November 2004. Munro & Associates. (In contrast. and that could not be adjusted to fit another model. MacDuffie interviews at Delphi. cited in Automotive News.